On November 2 the Democrats suffered a historic defeat, losing more than 60 seats in the House of Representatives. It was the worst midterm loss for the president’s party in the post–World War II era. Why did Democrats do so poorly? One week after the election, there is much we do not know, but political-science research and our own analyses of the election results suggest some provisional answers.

The "Fundamentals"

It was not a surprise that the Democrats lost their House majority. Much as presidential elections depend on certain national “fundamentals,” so do midterm elections. For instance, according to research by political scientist Gary Jacobson, about 80 percent of changes in House seats controlled by the president’s party can be explained using three factors: the number of seats the majority controls, the state of the economy, and the president’s approval rating. By the summer and early fall, it was clear that Democrats faced challenges on all three fronts.

First, Democrats were likely to lose seats simply because they held so many: more than 250 in the House, including 48 in districts that supported John McCain over Barack Obama in 2008. Second, voters tend to reward the president’s party when the economy is good and punish it when the economy is bad. Whether Obama and Congressional Democrats were at fault for continued economic weakness, the electorate held them responsible for it. Both factors also helped Republicans recruit good candidates, since such candidates are more likely to run when they have a better chance of winning. In 2008 only 28 percent of incumbent Democrats in potentially competitive seats faced Republicans with previous experience in elected office. In 2010 that number was 40 percent.

These powerful national forces created a significant swing toward the GOP across all districts. Compared to 2008, the Democrats lost vote share in 89 percent of districts. In the exit poll, almost every demographic group moved toward the GOP. Commentators were quick to single out groups, such as older voters and independents, that supposedly were crucial. But when there are consistent shifts among so many groups of voters, it seems clear that broad national-level forces were at work.

At the same time, most election forecasters still underestimated the Democratic losses. We don’t yet know why—that is, what other factors mattered that these forecasters did not or could not take into account. Perhaps the models didn’t fully account for the “enthusiasm gap” that manifested this year between Democratic and Republican voters or for the controversial nature of the Democratic legislative agenda. Perhaps the balance of fundraising and spending was important; turnout operations were key; or there were influential, but idiosyncratic, candidates. Time will tell.

The Districts

Even if we cannot yet explain the sheer extent of Democratic losses, we can get some purchase on why certain seats changed hands. The biggest part of the story—and by far the best predictor of where Democrats would lose—was the partisanship of individual districts. Democrats were most vulnerable in districts that were GOP-leaning or closely divided along partisan lines. We measured the partisanship of districts according to the percentage of district voters who chose Obama over McCain in 2008. Where Democratic incumbents won, Obama pulled 67 percent of the vote two years ago. But he received just 49 percent of the vote in the districts where incumbent Democrats lost in 2010. The 2008 presidential vote predicts wins and losses by Democratic incumbents with 85 percent accuracy.

Whatever benefits the Tea Party brought to the GOP, they did not appear to translate into major gains at the ballot box.

This powerful result provides some context for other interpretations of the election. For example, some commentators were impressed at the GOP’s gains in the Midwest and Northeast. But these two regions contain more than half the Democratic-held districts where Obama received less than 60 percent of the vote in 2008. Regional variation may still be important, but not as important as is sometimes implied.

The Tea Party

The 2010 election was widely seen as a test of the Tea Party’s influence. One could interpret the election as evidence of the movement’s popularity and mobilizing power, particularly because the GOP gains were larger than forecasting models anticipated.

With only one election since the Tea Party’s emergence, we cannot be sure that the Tea Party helped create the Republican “wave,” but we can compare the performance of candidates closely tied to the Tea Party to others who were not. Here, the record is mixed. Clearly, Tea Party groups endorsed some Senate candidates who proved to be liabilities—such as Sharon Angle in Nevada and Christine O’Donnell in Delaware—but other candidates, such as Marco Rubio in Florida and Pat Toomey in Pennsylvania, found success. Indeed, Angle and O’Donnell were the exception, not the rule. Many Tea Party–backed candidates in competitive districts and open-seat races had previous electoral experience and were not anti-establishment outsiders.

When we looked at the effect of Tea Party affiliation in House races, we found that Tea Party–backed Republicans did slightly better. But the difference was small—just a little over one percentage point—and could be the result of candidates in more pro-GOP districts making a strategic decision to affiliate with the Tea Party. Whatever benefits the Tea Party brought to the GOP, they did not appear to translate into major gains at the ballot box. But in the House at least, the movement did not damage Republican chances either.

The Money

Candidate fundraising and spending matter in congressional elections, particularly for challengers, who tend to be less well known than incumbents. We found that, controlling for other factors, in 2010 the balance of fundraising by opposing candidates was associated with the outcome. The more a candidate raised relative to his or her opponent, the more votes that candidate won.

In this election, however, the big story was not money raised by candidates, but by outside groups, particularly those supporting Republicans. Especially controversial were the groups that registered with the IRS instead of the Federal Election Commission to avoid disclosing their donors. However, our analysis found that the money spent by independent groups decided few House races. We predicted the outcome of each race including a measure of all money raised and a measure that subtracted the money spent by independent groups. Only two seats changed hands—and those actually shifted to the Republicans, not the Democrats, suggesting that the outside money may have hurt the GOP slightly.

We don’t place much credence in that result because it is difficult to measure the effects of campaign spending when donors tend to target the most competitive races. More important is simply that independent money did not seem to matter much. Why? One potential explanation is that outside spending was more balanced than media reports implied. The average independent expenditure for Democratic candidates (either for that candidate or against his or her opponent) was about $240,000. The average for Republican candidates was about $225,000. Even if we restrict the analysis to candidates receiving at least a million dollars in outside spending, the average for Republicans was about $1.6 million and for Democrats about $1.8 million.

Virtually every Democratic incumbent on the ballot supported at least one of the four bills advancing Obama’s agenda. That support was costly.

Moreover, the money was going into races that were already saturating the airwaves with advertisements, making it hard to overwhelm the opposition. The median amount of money raised by Democrats who faced more than $50,000 in independent spending was $1.7 million. The same number for Republican candidates was $1.3 million. This wasn’t David vs. Goliath. It was two Goliaths, beating each other to a pulp.

However, these conclusions come with two important caveats. First, we could not measure spending by the IRS organizations because it is not possible to link their spending to specific races. Second, we did not separate outside spending by parties from outside spending by other groups. The Republican Party was outspent by the Democratic Party in this election cycle, while the outside groups probably were tilted toward Republicans and helped make up the difference. But since our analysis only slightly under-predicts the actual number of seats won by the Republicans, fully accounting for both of these factors would affect a handful of seats at most—important, but hardly a game changer.
 

The Votes

Republican candidates and supporters were particularly critical of several policies of the Bush and Obama administrations. First was the Troubled Assets Relief Program, or TARP, which was enacted under President Bush in 2008 with the support of Obama (who was then the Democratic nominee). After Obama took office, House Democrats passed the stimulus package, the cap-and-trade bill, and, most (in)famously, health care reform (the first and third were enacted into law, while the second died in the Senate). Voters sometimes hold representatives accountable for their roll call votes, punishing those whose voting records are ideologically out of step with their districts. Did this happen?

We counted the number of these measures supported by each Democratic incumbent and then estimated the effect of this support on their election conditional on the partisanship of the member’s district (controlling for other factors). The simple answer: these roll call votes mattered. A lot. A Democratic incumbent in the average district represented by Democratic incumbents actually lost about two-thirds of a percentage point for every yes vote. Democrats in the least Democratic districts, such as Chet Edwards of Texas or Gene Taylor of Mississippi, lost about 4 percent for every yes vote. Does that mean poor Chet lost sixteen points on roll call votes alone? No, because he wasn’t a big supporter of Obama’s agenda. But he did vote for both TARP and the stimulus. In fact, virtually every Democratic incumbent on the ballot supported at least one of these four bills. That support was costly.

Using our statistical model, we simulated what would have happened if every losing Democratic incumbent had not supported any of these bills. The results are dramatic: the Democrats would have gained back 32 seats, enough to retain control of the House. The simulation involves some uncertainty, so we can’t be sure that voting differently on these issues would have guaranteed a retained majority, but clearly the Democrats would have lost significantly fewer seats.

Of course, there were almost no Democratic incumbents running for reelection who voted against all four bills. If we just make the vulnerable Democrats vote no on one more of the four bills than they actually did, the gain for Democrats is only twelve seats, not 32.

The 2010 election leaves Obama with less power to promote his economic agenda but all the accountability.

Our simulations only estimate the effect of different voting records—the actual counterfactuals are more complex. For example, a lack of Democratic support on TARP or the stimulus could have killed those measures, potentially leaving the economy in worse shape and hurting the Democratic Party’s chances in 2010 even more than these measures’ passage. We are not claiming that the Democrats shouldn’t have voted for these bills, only that Democrats took some tough votes and paid an electoral cost for them, especially in marginal districts.
 

The Next Two Years

The 2010 election put the president and Congress in a familiar position: divided government. What does that mean? The weeks and months preceding the election produced a lot of counterintuitive thinking, with suggestions that divided government would somehow make government more productive or help Obama get reelected. These notions are too clever by half. Divided government is not good for presidents. Political scientists have found that much less of the legislative agenda gets enacted under divided government, in part because presidents are more likely to oppose legislation produced by Congress.

The shift to divided control might be less consequential for Obama if voters considered the president and Congress equally to blame. But that’s not likely. In an analysis of over 40 years of presidential elections, political scientists Richard Nadeau and Michael Lewis-Beck found that voters rewarded the president's party when times were good and punished it when times were bad regardless of whether government was unified or divided. Thus, the 2010 election leaves Obama with less power to promote his economic agenda but all the accountability. If the economy is strong in 2012, he is likely to be reelected; if it is weak, he is likely to lose.

The White House can take solace only in the challenges that divided government poses for the opposition. The Republican Party has embraced many demands by conservative activists both within and outside the Tea Party, including the repeal of health care reform and sharp reductions in taxes and federal spending. But they will also find their legislative ambitions frustrated under divided government, which often requires compromises that may alienate their uncompromising base. And the GOP’s ambitions to retake the executive branch are largely at the mercy of the economy. In this respect, the president and his new congressional “partners” are equal.